[mvapich] Announcing the release of MVAPICH 0.9.5 + sample performance numbers

[Dhabaleswar Panda]

You are receiving this e-mail because of: 

  1) being a member of more than 200 organizations (in 26 countries)
  who have downloaded MVAPICH/MVAPICH2 (MPI over InfiniBand) from the
  OSU web site

  and/or 

  2) your interest in the MVAPICH/MVAPICH2 package, developed in the
  Network-Based Computing Laboratory at the Ohio State University.

The MVAPICH team is pleased to announce the release of MVAPICH 0.9.5
for multiple platforms (EM64T, G5, IA-32, IA-64, and Opteron) and
network interfaces (PCI-X and PCI-Express-including the new mem-free
cards).

MVAPICH 0.9.5 is being distributed as a single integrated package
(with the latest MPICH 1.2.6 and MVICH). It can be downloaded with a
`single click' and installed. It is available under BSD license.

This new release has the following features:

      - multi-rail support (multiple adapters per node and 
                 multiple ports per adapter) 

      - optimized intra-node shared memory support 
                 (both for bus-based and NUMA-based systems)  

      - enhanced MPI broadcast support with IBA hardware-based
                 multicast

      - flexible mechanisms for minimizing memory resource 
                 usage on large scale clusters 

      - support for TotalView debugger 

      - optimized and tuned for the above platforms and different 
                 network interfaces (PCI-X and PCI-Express)

      - single code base for all of the above platforms

Other features of this release include:

- Excellent performance: MVAPICH 0.9.5 with multi-rail (1-NIC, 2-port)
  delivers 4.0 microsec latency, up to 1498 MB/sec unidirectional
  bandwidth, and up to 2704 MB/sec bidirectional bandwidth on EM64T 
  system with PCI-Express. Detailed performance numbers for other 
  platforms are available on the project's web page. 

- An enhanced and detailed `User and Tuning Guide' to assist users: 

       - to install this package on different platforms 
            with different options

       - to vary different parameters of the MPI installation to 
            extract maximum performance and achieve scalability, 
            especially on large-scale systems.

You are welcome to download the MVAPICH 0.9.5 package and access
relevant information from the following URL:

http://nowlab.cis.ohio-state.edu/projects/mpi-iba/

Since the 0.9.4 release, we have introduced a set of patches based on
user feedbacks. If you plan to continue using 0.9.4 for some more
time, we strongly encourage you to download and apply these patches to
your current installation.

Our upcoming releases include:

    - an OpenIB Gen2 version of MVAPICH 0.9.5

    - MVAPICH2 0.6.5 with uDAPL support to run on different networks
      with uDAPL interface

All feedbacks, including bug reports and hints for performance tuning,
are welcome. Please send an e-mail to mvapich-help at cse.ohio-state.edu.

Thanks, 

MVAPICH Team at OSU/NBCL 

----------

PS: If you would like to be removed from this mailing list, please end
an e-mail to mvapich_request at cse.ohio-state.edu.